Workers in the field of high-speed document processing, such as in the sorting of bank checks and like financial instruments, know that the art requires the use of machines and systems capable of moving and processing very large volumes of documents at rates of thousands of documents per minute, while performing multiple and interrelated operations upon each document as it travels through such machinery. Such operations might include, but are not limited to, printing upon the documents, reading data previously encoded thereon by a variety of processes, recording an archival image of the document by photographic or electronic-imaging techniques, and other processes and manipulations.
The "Doubles" Problem:
Workers understand that, while processing such large volumes of documents, it is vital that each individual document be transported and processed singly, and that documents remain in the order and sequence in which they were processed by the machine. To attain the rates of document processing required, the documents are fed and separated from one another by machinery, which is extensively designed and engineered to ensure that documents are fed one at a time ("Singly") with a very high degree of reliability. Should two or more documents be accidentally fed and processed together, extensive manual effort and time are required to track down this error among the many thousands of documents which the machine may process within a very brief time. For this reason, the most extreme measures are implemented to ensure that the document feeding and separating measures always feed documents one at a time, no matter what their condition.
Nonetheless, there are occasional unavoidable circumstances where the machinery will feed more than one document at a time. Examples are documents which are stapled or glued together, documents which adhere to one another due to ink or other surface treatments, or documents which are attached one another by mutual tears or folds. Such cases are known in the art as "double-documents" or simply "doubles". Human operators for such sorting and processing machinery are aware that "doubles" are a costly and time-consuming event, and guard against them as far as possible; still, the sheer volume of documents means that a "double" will occur from time to time.
For this reason, workers find that the machinery itself must contain a reliable device for separating and detecting "doubles" as soon as possible after they have been fed; preferably before much processing is performed on them. In this way, the operator may be warned of the presence of a "double" before it can cause a disruption to the normal flow of work (e.g., and remove it).
We have contemplated different techniques for reliably sensing and reporting a double-document. Such techniques must take account of the widely varying characteristics of the documents (e.g., thickness, density, opacity, etc.), as well as the increasing document speeds which are the result of continuing efforts to increase the processing rate. Theoretically, "doubles" might be sensed optically, mechanically or electronically--as noted below:
Optical Sensing:
By shining a beam of light through the document and measuring how much of the beam passes through to impinge on a sensor, the additional thickness of a second document should produce a measurable change in signal.
This technique, while practical in principle, tends to perform poorly in service. The wide range of characteristics of the documents being fed, especially as regards opacity and thickness, renders such a technique difficult to implement in practice. Because such a system must tend to operate in a "fail-safe" mode, it has to lead to a high incidence of false "doubles". Such a false report is almost as disruptive as a real "double" would be.
Additionally, optical sensors are very susceptible to failure due to the high levels of dust and debris found around document processing machinery.
Mechanical Sensing:
By passing the document between a known reference point and some moving effector, such as a stylus or roller, the thickness of the document may be measured by means of one of a variety of sensors. The additional thickness of a second document should be measurable.
Once again, the wide range of characteristics for documents fed make this a poor system. The thickest documents may well be more than twice the thinnest, causing a high incidence of "false doubles". Additionally, the sensors required to detect mechanical variations of this order are sensitive and costly, and require skilled and time-consuming calibration to give a reliable result.
Electronic sensing, relying on the variation of a parameter such as reluctance or permeability to detect the presence of documents. PA1 No gradual changing pressure drop with the cyclone as with a barrier type filter. PA1 Quick and easy bunker clean out; simply draw the collected dust out of the bunker with a standard vacuum cleaner.
Again, the range of document characteristics render such techniques less than successful, also they require the use of costly, custom-sensing elements.
Rather, we settled on a vacuumatic separating/sensing technique; and found it to give high reliability regardless of the nature and condition of the documents. This invention seeks to teach improvements in such techniques to enhance reliability, serviceability and whole-life cost.
Basic Vacuumatic System (FIGS. 1, 1A):
FIGS. 1, 1A, 2 show a basic, simplified version of a vacuum-separation and sensing system of the type we first favored. Here, it will be understood that the documents to be sensed are transported in a vertical position by transport means such as belts, pulleys and the like (not shown, but well understood in the art). The documents d are constrained to pass through a vacuum-separation manifold M which encloses the lower longitudinal edge of the document as it passes.
This manifold incorporates two vacuum ports V1 and V2, each disposed on a respective side of the document. The two ports are connected to a common plenum chamber P, which is kept at negative pressure relative to the surrounding atmosphere by vacuum blower means B, (or the like) connected to plenum P by hose means H.
Connected to a port Q provided in the wall of hose means H is a differential pressure switch S1 which compares the pressure within the hose to the ambient atmospheric pressure.
When no document is present, both vacuum ports V1 and V2 are open and unobscured, and air may freely flow into them under the influence of blower means B. The pressure differential between the inside of hose H and the surrounding atmosphere is "LOW" (.DELTA.P.sub.L).
When a single document passes through manifold M, it will be pulled towards one or other of the two vacuum ports V1-V2 by the suction applied from blower B; the document will tend to close off whichever of the ports it is first drawn to. The other port will remain open and unobstructed. The document will cause some reduction in the airflow through ports V and there will be a "Moderate" pressure differential (.DELTA.P.sub.M) between the inside of hose H and the surrounding atmosphere.
When two or more documents pass through manifold M side-by-side (or just overlapping), they will tend to be separated by the suction applied from blower B and each will tend to be drawn to an adjacent vacuum port V. When one port is closed and blocked by a document, the suction at the other port will be increased by virtue of the restriction of the airflow, so this port will tend to drawin the second document even more strongly. When both ports are thus closed and blocked, airflow is very quickly reduced to almost nil very quickly and the differential pressure between the inside of hose H and the surrounding atmosphere will very quickly rise to the highest level of vacuum (.DELTA.P.sub.h) which blower B is capable of sustaining at this point.
It will thus be seen that, by monitoring the pressure differential between the inside of hose H and the surrounding atmosphere, an indication of the presence of more than one document in manifold M may be obtained which is more or less independent of any physical characteristic of the documents (such as opacity, thickness, color and so on) and is also independent of the number of documents present. By selecting a threshold of pressure differential B for switch S1 which corresponds to "both ports V covered" (e.g., .DELTA.P.sub.h), such an arrangement can automatically indicate "more than one document", regardless of the actual number of documents involved, and regardless of their individual characteristics.
The action of switch S1 is converted to an electrical signal, which is processed by signal-conditioning circuitry (not shown, but familiar to workers in the art) and provides to the controlling systems of the (check-sorting) machine an indication that a "double" has been detected. The controlling systems can then direct the suspected "double" to a holding area of the machine, without further processing, and alert the machine operator, who may investigate the item manually to correct or otherwise resolve the "double".
Since such a "doubles-detect" arrangement was first contemplated, there has been significant progress in the design of check sorting machines. Document speeds and feed rates have increased, and the types and quality of documents handled have expanded beyond any expectation. Additionally, expectations are now greater; e.g., as to convenience of operation, cleanliness, hygiene, and safety. Modifications have to be made to meet these needs. Among these conditions are the following:
Re Separation/Sense Time:
Increasing document speeds have reduced the time available for a "doubles-detect" system to operate on a passing document and determine whether it is a "double". As an example, the Unisys DP1800 check sorting machine operates at a nominal track speed of 300 inches per second (ips, or 7.62 meters per second), and may operate with documents with a minimum length of 5.75 inches (11.4 centimeters). For such a document, the time available to operate on a document (e.g., to separate) is 5.75/300 seconds, or about 19 milliseconds. Future developments are likely to increase document speed to as much as 400 ips (10.1 meters per second), with a corresponding reduction in time available for a sensor to make its determination. To allow a system to operate adequately within such reduced time periods, larger and more powerful blowers (B) have to be employed. A blower for the DP1800 product, for example, would be rated to flow 30 cubic feet of air per minute and provide a maximum vacuum of 30 inches water gauge. These high airflows and vacuums would be required to ensure that the "double" is separated within the manifold M as quickly and securely as possible, even when the documents consist of heavier paper stock with higher resistance to "bending".
Re Dust:
Increasing document speeds and a wider range of document types lead to more dust and debris being generated in the machine. This material may consist of paper fragments and dust, generated by the friction of document-driving elements or from the cut and sheared edges of the paper itself, as well as rubber and plastic particles shed from the driving elements (e.g., rolls, belts and the like, as well understood by workers in the art).
Paper handling business machines (e.g., Unisys check processors) employ vacuum systems to transport or detect documents or for other functions. The vacuum is generated by vacuum pumps or blowers. These pumps/blowers require filtration of the air they move to protect their internal moving components from damage from dust/dirt in the air. Additionally, any exhaust air must be filtered to prevent contamination of the customer's office environment.
Since paper handling machinery usually generate lots of paper dust, the pump/blower air filters tend to quickly fill with dust.
Typically the air filtration systems used are "barrier type" i.e., fiberglass or porous filter paper of some type. These require frequent field service maintenance for cleaning or replacement. In a high volume site for a Unisys DP1800 document processor, these filters typically require replacement twice a week. This frequent servicing by skilled field engineers adds substantially to the maintenance cost of this type business machine.
One advantage of a "cyclone" paper dust collector is that it can contain relatively large amounts of dust in its bunker, and so reduce the required frequency of maintenance. In a DP1800 document processor for example, using a cyclone filter/blower embodiment can reduce frequency of service from twice a week to once every 3 months (or 1:24 ratio); and there are other advantages, such as:
Now, such dust/debris will naturally be drawn into the manifold M of a doubles-detect system under the action of the vacuum generated by blower B, and it may collect within the system. There, it may clog pipes and hoses, such as the connection to the pressure switch S1, or it may build up inside the blower B to the point where performance is reduced, requiring extensive maintenance and reducing machine up-time. Finally, such material will be (mostly) ejected from the system in the exhaust of blower B, into the surrounding atmosphere, where it creates an unsightly and unhealthy environment for attendants. It can also constitute a fire hazard if allowed to accumulate, both inside the machine and in the surrounding environment.
.DELTA.P as Mini-Pulses:
Increasing speeds, and the resulting need for increasing vacuum, have also led to subtle changes in the way that a system must function and provide sensing output. The changes in pressure detected by the pressure switch S1 have become less of a mass-air-flow phenomenon (as they were at lower speeds and lower airflows) and more of a "pulse" phenomenon. Where document speeds are slower, pressures would rise and fall (in response to the states of vacuum ports V) relatively slowly and evenly throughout the system. With much higher speeds and airflows, and much shorter transitions at vacuum ports V, pressure changes now move through the system as a "pulse" of reduced pressure, entrained in a high-speed column of air moving through the system. While this is not a problem in and of itself (since the pressure switch S1 can still detect and respond to such "pulses" in the same way as if they were a more general reduction in pressure throughout the entire system), precautions have to be taken to prevent minor, spurious pulses ("mini-pulses", or transient spikes) of changing pressure from being generated and producing false results at the pressure switch. To this end, sensing port Q is moved further down hose H from manifold M to provide an effective column of air within hose H between manifold M and sensing port Q. The mass and volume of this column can act as a dynamic damper for pressure variations travelling there along, and can attenuate the magnitude of such pulses as they travel from manifold M to sensing port Q. In this way, the impact of such pulses at pressure switch S1 may be reduced, --though not entirely eliminated.
Similar problems from "mini-pulses" can also be caused by the documents themselves as they travel through manifold M. As document speeds increase, aerodynamic effects become more and more significant. A document's leading edge may "hunt" from side to side; also the entire document may assume one of several conditions, such as an undulation from side to side along its length, or a tendency to travel at an angle to the direction it is being driven in. These conditions may, in turn, lead to unexpected results at manifold M, where a single document may rapidly obscure first one vacuum port V, then the other, setting up a series of high-frequency pulses in the airflow, in Manifold M and in the various hoses connecting it to blower B. Thus, pressure switch S1 must be carefully designed and tested to ensure that such mini-pulses do not cause spurious signals. Also, more stringent measures should be taken to selectively damp the airflow to filter out and negate such mini-pulses.
Adding Filtration (FIG. 2):
FIG. 2 shows modifications in the FIG. 1 arrangement for addressing some or all of foregoing concerns. The system is altered by addition of a mechanical air filter F in hose H, and provision for adjusting the airflow through hose H is made by adding a variable orifice R at the entrance to air filter F. The air filter serves to separate dust and debris from the air stream before it enters blower B, and to prevent it from clogging the blower and/or being expelled into the surrounding environment. Variable orifice R allows the air flow (and therefore the system differential pressures), to be calibrated to a known standard, which is typically measured by applying a vacuum gauge (not shown, but well known in the art) to a test port T provided in the body of the air filter housing.
While these measures address the identified system problems, they bring problems of their own and generate new system problems. The air filter, for instance, will soon become clogged with dust and debris, thus tending to restrict airflow and alter the differential pressures within the system. As this restriction increases with the buildup of debris, the system will tend to miss doubles, since the vacuum at manifold M would be reduced. Thus, a system to warn the operator of excessive buildup of filtered material is desirable--e.g., consisting of a pressure switch S2 which measures the differential pressure across the filter and warns the operator when it reaches a predetermined level, indicating that the filter is excessively clogged. As speeds increase, the replacement period for filter elements will decrease in proportion, until, in some systems, these filters will need to be replaced every few days to maintain consistent system performance.
Also, variations in filter elements etc., will typically make it necessary to check and adjust system pressures every time the filter is changed. One can do this with variable orifice R, allowing an attendant to adjust system pressures to a known standard. This practice, while improving system performance, and maximizing filter replacement intervals, adds considerably to service time and cost.
All the foregoing conditions combine to produce a set of requirements far more stringent than were originally conceived. Since air flow rates are far higher, the system must be designed with a minimum of restrictions which might reduce the flow or produce undesired pressure effects. Variations in airflow (and therefore pressure) must be kept to a minimum over the long term, to maximize the thresholds defining a "double" and to minimize the incidence of false signals. And the system must accommodate a large and continuous supply of dirt and debris without impacting its function and (preferably) without ejecting a lot such material into a customer's environment.
This invention addresses these and related problems; e.g., teaching a doubles-separation/doubles-detect arrangement using vacuumatic means, teaching such with opposed vacuum-ports and associated pressure-sensing means to signal the presence of a single document or overlapped documents; preferably by locating such sensing means sufficiently remote from such ports to provide a damping-column adequate to attenuate, or mask-out, minor pressure variations; by teaching air-filter means and related variable orifice means and filter-pressure-sensing means to adequately filter-out contaminants entrained in the line from such ports to the vacuum source, while also allowing one to be aware of excess across-filter blockage, and to "re-tune" the system pressures once a filter element is installed/replaced.
And, beyond the foregoing, it is an object hereof to teach such a system wherein the vacuum manifold and sensing means are integrated into a single unit of minimal and controlled variability, while the supporting systems are so designed as to provide optimum airflow over long periods with minimal maintenance. The taught arrangement also includes a function whereby the sucking vacuum is kept essentially constant, and whereby dirt and debris are automatically extracted from the airflow as a function of its operation, but without use of a barrier filter and with minimal impact upon normal operation, and means whereby such foreign matter may be accumulated over long periods and purged from the system without impact upon its normal operation.
The methods and means discussed herein will be generally understood as constructed and operating as presently known in the art, except where otherwise specified; and with all materials, methods, devices and apparatus herein understood as being implemented by known expedients according to present good practice.